Development of the next-generation GPU-based Monte Carlo simulation platform for radiation-induced DNA damage calculations

开发下一代基于 GPU 的蒙特卡罗模拟平台，用于辐射引起的 DNA 损伤计算

• 批准号: 10203527
• 负责人: Yujie Chi
• 金额: $ 44.66万
• 依托单位: UNIVERSITY OF TEXAS ARLINGTON
• 依托单位国家: 美国
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-05-01 至 2025-04-30
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10203527
• 关键词: Address; Advanced Development; Affect; Biological; Biological Process; Biomedical Research; Cell Cycle; Chemicals; Clinical; Communities; Consumption; DNA; Data; Development; Diagnostic Imaging; Dose; Education; Educational process of instructing; Electrons; Ensure; Exhibits; Exposure to; G1 Phase; Geometry; Goals; Hispanic-serving Institution; Human; Industry; Ionizing radiation; Knowledge; Learning; Malignant Neoplasms; Mammography; Medical Imaging; Medicine; Metaphase; Methods; Microscopic; Military Personnel; Mission; Modality; Modeling; Modern Medicine; Monte Carlo Method; Organism; Outcome; Oxygen; Phase; Physics; Play; Polymers; Positron-Emission Tomography; Probability; Problem Solving; Process; Radiation; Radiation exposure; Radiation therapy; Radiosurgery; Research; Role; Shapes; Structure; Students; System; Technology; Testing; Therapeutic; Thoracic Radiography; Time; Uncertainty; Water; base; carcinogenicity; cost effective; design; experience; genotoxicity; imaging modality; improved; indexing; ionization; model development; molecular dynamics; multi-scale modeling; next generation; novel; open source; parallel computer; particle; repaired; simulation; spatiotemporal; success; tomography; tool; undergraduate student; user-friendly; web interface

Project Summary Ionizing radiation (IR) is a critical component of modern medicine. When IR penetrates through the organism, it could depart its energy to the medium mainly through ionization and excitation. The energy departure of IR is medium composition dependent, and hence it is used to ‘see’ the inner structure of the human beings, enabling the application of IR in the medical imaging of mammography, chest x-rays, computational tomography, positron emission tomography, etc. IR can also damage the structure and/or affect the function of the organism and hence it is applied to treat cancer in the form of radiosurgery and radiotherapy. Meanwhile, IR is found to be genotoxic and carcinogenic, calling the non-ending effort to understand the fundamental effects. Advanced cellular radiobiological study exhibited that the damage of deoxyribonucleic acid (DNA) plays a pivotal role towards the determination of the final biological or even clinical outcome after exposure to IR. It is hypothesized that when IR interacts with DNA, it could damage DNA in picoseconds by the primary and secondary IR particles and in microseconds by subsequently generated radiation radicals. It is then essential to understand how IR produces this initial damage under various radiation conditions. Microscopic Monte Carlo (MC) simulation such as Geant4-DNA, capable of computing this damaging process, has been playing an important role in the quantitative hypothesis-test. However, there are several issues in the state-of-the-art MC tools, making it hard to meet the increasing demanding for advanced applications. These include the low efficiency in dealing with the ‘many-body’ problem, the relatively large uncertainty in the final computing results, the lack of support for the entire cell cycle and the limited-access/user-unfriendly designs, etc. In this project, we propose to solve the above issues by developing a next-generation MC simulation tool for IR induced DNA damage computation through the novel implementations of graphical processing units (GPUs) parallel computing, the molecular dynamics/first principles based computation, the new DNA model development based on the extrusion model and polymer physics, and the open-source release with user-friendly interface. Upon success, the developed system is expected to serve as a next-generation simulation platform for the calculation of the initial DNA damage caused by IR, which can become a profound first-step towards a successful accomplishment of the “bottom-up” multi-scale modeling for the entire radiobiological process, making a significant impact in radiation medicine.

项目摘要 电离辐射(IR)是现代医学的重要组成部分。当红外线穿透生物体时，它 可以主要通过电离和激发将其能量转移到介质中。IR的能量起点是 媒介成分依赖的，因此它被用来‘看到’人类的内部结构，使 红外在乳腺摄影、胸部X光、计算机体层摄影、正电子医学成像中的应用 红外线也会破坏结构和/或影响生物体的功能，因此 它以放射外科和放射治疗的形式应用于治疗癌症。同时，IR被发现具有遗传毒性 和致癌，称为无休止的努力，以了解根本的影响。 先进的细胞放射生物学研究表明，脱氧核糖核酸(DNA)的损伤起着关键作用 在确定暴露于IR后的最终生物学甚至临床结果方面起着重要作用。它是 假设当IR与DNA相互作用时，它可以在皮秒内通过初级和 二次红外粒子和随后产生的辐射自由基在微秒内。因此，至关重要的是 了解红外线是如何在各种辐射条件下产生这种初始损害的。微观蒙特卡罗 (MC)模拟，如Geant4-DNA，能够计算这种破坏过程，一直扮演着 在定量假设检验中的重要作用。然而，在最先进的MC中有几个问题 工具，使其难以满足对高级应用程序日益增长的需求。其中包括低点 处理多体问题的效率，最终计算结果的相对较大的不确定性， 缺乏对整个细胞周期的支持，以及访问受限/用户不友好的设计等。 在这个项目中，我们提出了通过开发下一代面向信息检索的MC仿真工具来解决上述问题 通过图形处理单元(GPU)的新实现进行DNA损伤计算 并行计算、基于分子动力学/第一性原理的计算、新DNA模型的发展 基于挤出模型和聚合物物理，并开放源码发布，界面友好。 一旦成功，开发的系统有望成为下一代的仿真平台 计算红外引起的初始DNA损伤，这可以成为迈向成功的深刻的第一步 完成了对整个放射生物学过程的自下而上的多尺度建模，使 对放射医学产生重大影响。